An advance in adoptive T cell therapy is the ability to endow patient's T cells with reactivity for tumor cell surface molecules through the introduction of genes that encode synthetic chimeric antigen receptors (CARs). Studies of small numbers of patients with advanced B cell malignancies have demonstrated potent antitumor effects of T cells that express a CD19-specific CAR in a subset of patients, suggesting this approach could provide a major advance in therapy. However, not all patients respond and the duration of response remains uncertain. Prior studies have not rigorously defined the phenotypic composition or frequency of CAR T cells administered to patients, which has resulted in variations in potency and in vivo persistence, and could explain the lack of efficacy in many patients. Our group has focused on elucidating cell intrinsic properties of human T cells that provide for reproducible in vivo behavior after genetic modification and adoptive transfer, and on optimizing the design of CARs for tumor recognition and safety. We initiated the first clinical tril of CD19 CAR- T cell therapy in which the T cell products administered to every patient were formulated in a defined optimized composition. The initial results have revealed profound tumor regressions in patients with advanced chemotherapy refractory NHL after infusion of small doses of CD19 CAR-T cells. This proposal will build on these accomplishments and develop CD19 CAR T cell therapy into a reproducible, broadly effective and safe therapy for B cell malignancies. The specific aims are: Aim 1. To evaluate the safety, antitumor efficacy, and mechanisms of tumor escape after adoptive transfer of CD19 CAR-T cells administered in a defined cell product composition in patients with refractory B cell lymphoma or leukemia. Aim 2. To evaluate the safety and durability of antitumor responses of CD19 CAR-T cells derived from TM cells for therapy of ALL and CLL after HLA matched related or unrelated allogeneic HCT. Aim 3: To perform a phase I clinical trial to determine if adoptively transferred CD19 CAR-T cells that co-express a truncated human EGFR can be deleted in vivo by infusion of the anti-EGFR mAb, Erbitux.